Update of /project/cells/cvsroot/cells/tutorial In directory clnet:/tmp/cvs-serv2216/tutorial

Modified Files: 01c-cascade.lisp Log Message: evolving geometry; refinement of test case 01c-cascade

--- /project/cells/cvsroot/cells/tutorial/01c-cascade.lisp 2006/05/30 02:47:45 1.1 +++ /project/cells/cvsroot/cells/tutorial/01c-cascade.lisp 2006/06/05 00:01:22 1.2 @@ -1,31 +1,94 @@ -#| Now we have automatic state management (including change propagation) +#| + +Now we have automatic state management (including change propagation) outside the Cells model as well as in. Now lets look at cascading change by adding another level of computation, so A->B->C.

-[Actually, I see I need to make this a little deeper, since area has -a direct dependency on width. Not tonight. :)] +In this case: len->area->brightness +Also: len->width->area->brightness + +That leads to some complications I will discuss, but no assertions here +enforce correct behavior in re those complications. Soon. :)

|#

(defpackage #:tu-depth (:use :cl :cells)) (in-package #:tu-depth)

+(defmacro start-finish (key rule) + `(progn + (print (list :start ,key)) + (prog1 + (progn ,rule) + (print (list :finish ,key)))))

(defmodel rectangle () - ((area :initarg :area :accessor area - :initform (c? (print :compue-area) - (* (len self)(width self)))) + ((lumens :initform 1000000 :reader lumens) (len :initarg :len :accessor len - :initform (c? (print :compute-len) - (* 2 (width self)))) + :initform (c? (start-finish :len + (* 2 (width self))))) + (area :initarg :area :accessor area + :initform (c? (start-finish :area + (* (len self)(width self))))) (width :initarg :width :accessor width - :initform (c? (print :compute-width) - (floor (len self) 2))))) + :initform (c? (start-finish :width + (floor (len self) 2)))) + (brightness :reader brightness + :initform (c? (start-finish :brightness + (/ (^lumens) (^area))))) + ))

#+test -(let ((r (make-instance 'rectangle :len (c-in 42)))) - (cells::ct-assert (eql 21 (width r))) - (cells::ct-assert (eql (* 21 42) (area r))) +(let ((r (make-instance 'rectangle :len (c-in 100)))) + (cells::ct-assert (eql 50 (width r))) + (cells::ct-assert (eql 5000 (area r))) + (cells::ct-assert (eql 200 (brightness r))) (cells::ct-assert (= 1000 (setf (len r) 1000))) - (cells::ct-assert (eql 500000 (area r)))) + (cells::ct-assert (eql 500000 (area r))) + (cells::ct-assert (eql 2 (brightness r)))) + +#| --- discussion ---------------------------- + +The output in Cells is: + +(:START :AREA) +(:START :WIDTH) +(:finish :WIDTH) +(:finish :AREA) +(:START :BRIGHTNESS) +(:finish :BRIGHTNESS) +(CELTK::ATTEMPTING (EQL 50 (WIDTH R))) +(CELTK::ATTEMPTING (EQL 5000 (AREA R))) +(CELTK::ATTEMPTING (EQL 200 (BRIGHTNESS R))) +(CELTK::ATTEMPTING (= 1000 (SETF (LEN R) 1000))) +0> c-propagate-to-users > notifying users of | [i :=[24]LEN/#<RECTANGLE>] | (AREA WIDTH) + +Notice here that the LEN cell is about to tell both the width and area to recalculate, +since area depends (of course) on len and (rather artificially) width also derives +from LEN. + +ie, This example has accidentally deviated into more complexity than intended. But we are +approaching these issues anyay, so I will leave it for now. We can always break it up +later. + +Let's continue: + +(:START :WIDTH) +(:finish :WIDTH) +(:START :AREA) +(:finish :AREA) + +Fine, now here comes the challenge. Width is also going to tell area to recalculate: + +0> c-propagate-to-users > notifying users of | [? :<vld>=[24]WIDTH/#<RECTANGLE>] | (AREA) +0> c-propagate-to-users > notifying users of | [? :<vld>=[24]AREA/#<RECTANGLE>] | (BRIGHTNESS) + +Correct: Area does not actually run its rule since it already did so when notified by LEN, + but it does propagate to brightness. + +(:START :BRIGHTNESS) +(:finish :BRIGHTNESS) +(CELTK::ATTEMPTING (EQL 500000 (AREA R))) +(CELTK::ATTEMPTING (EQL 2 (BRIGHTNESS R)))

+|# \ No newline at end of file